SUCK - Common Questions Not Answered Elsewhere

Last Updated - 2 Feb 99

For Inn related questions see the Inn/Suck FAQ, maintained by Philip Warner

I intend to use the page to answer questions that I get that aren't necessarily suck questions, but related to suck, such as common INN problems. Comments, corrections, suggestions, are always welcome.

1. Why am I getting '430 no such article' Errors?

Here is a somewhat long explanation about how suck gets the articles, and some possible explanations: Here's the way suck gets the message-ids to download:

Pseudo code of get_message_index function
 
for each GROUP {
	GROUP command
	XHDR msgid msgnr- 
}

The GROUP command tells the NNTP server which group I am interested in, then the XHDR commands says to send me all msgids for the range specified in sucknewsrc (last high msgnr usually). The remote NNRP server typically gets this information from the HISTORY or XOVER database it maintains. It has NOT yet looked at the actual articles on the spool directory. This gives me the raw, undeduped list of msg-ids that I have to worry about. Then, I take the list of msgids, dedupe em against itself (so I don't dl the same article twice) and the history database (if requested). Once the deduping is done, I then do:

pseudo code for get_messages function

for each MSGID {
	article MSGID
}

Of course, this gets more complicated, based on killfiles, etc. This is the point when the NNRP server goes out to get the article from the spool directory. Now, if for any reason, and there are many (badly deleted articles, short retain date, currently running news.daily), the article is deleted from the spool directory, but not from the database of msgids, you'll see the error message noted. If you get these on a daily basis, it can mean that your server is not keeping up with its incoming/outgoing articles. Try using the -n option to suck. This forces suck to use the article number vice MsgID to retrieve the article. If this works, it definitely means the problem is on the remote end. The code that I use to do this processing hasn't changed in a few releases. That's why I don't think its a bug in suck. HOWEVER, there are a couple of ways to very quickly test that theory. If suck gives you a 430 error code, take that MSGID, and put it in "suckothermsgs", this will force suck to try and download it again. Or manually connect to the NNTP server:

telnet remote.server 119
mode reader (if you need the -M command) 
article  (include the <>s)
quit

If either one of these does succeed in getting the message, then there is a very real chance that suck has a very nasty hidden bug somewhere. Let me know. I will need the debug output from a session (use the -D option).

2. I keep getting "article unavailable" when talking to a SuperNews server, what do I do?

Answer - Use the -n option. For some reason, Supernews doesn't like it when Suck requests articles using the MsgId.

Comments? Mail me at bobyetman@worldnet.att.net